Review series

Inflammation and insulin resistance

Steven E. Shoelson, Jongsoon Lee, and Allison B. Goldfine

Joslin Diabetes Center and Department of Medicine, Harvard Medical School, Boston, Massachusetts, USA.

Over a hundred years ago, high doses of salicylates were shown to lower glucose levels in diabetic patients. This should
have been an important clue to link inflammation to the pathogenesis of type 2 diabetes (T2D), but the antihypergly-
cemic and antiinflammatory effects of salicylates were not connected to the pathogenesis of insulin resistance until
recently. Together with the discovery of an important role for tissue macrophages, these new findings are helping to
reshape thinking about how obesity increases the risk for developing T2D and the metabolic syndrome. The evolv-
ing concept of insulin resistance and T2D as having immunological components and an improving picture of how
inflammation modulates metabolism provide new opportunities for using antiinflammatory strategies to correct the
metabolic consequences of excess adiposity.

Historical perspectives on the link between inflammation
and insulin resistance
Clues to the involvement of inflammation in diabetes date back
to more than a century ago, when high doses of sodium salicylate
(5.0–7.5 g/d) were first demonstrated to diminish glycosuria in
diabetic patients having “the milder form of the disease,” presum-
ably type 2 diabetes (T2D) (1–3). In 1876 Ebstein concluded that
sodium salicylate could make the symptoms of diabetes mellitus
totally disappear (1, 3). Similarly, in 1901 Williamson found that
“sodium salicylate had a definite influence in greatly diminish-
ing the sugar excretion” (2). The effect was rediscovered in 1957
when an insulin-treated diabetic, given high-dose aspirin to treat
the arthritis associated with rheumatic fever, no longer required
daily insulin injections (4). Fasting and postchallenge glucose
concentrations were nearly normal, despite the discontinuation
of insulin and treatment with aspirin alone. Upon resolution of
the joint symptoms, the aspirin was discontinued, and a repeat
glucose tolerance test was grossly abnormal. Intrigued by these
findings, Reid and colleagues prospectively studied 7 additional
patients, 4 the “overweight mild type” and three “lean more severe
diabetics” (4). Over a 2-week course of high-dose aspirin (5.0–8.0
g/d), fasting blood glucose levels fell from an average of more than
190 mg/dl before treatment to 92 mg/dl; every patient responded.
Additional trials showed equivalent efficacy (5, 6). Mechanistic
studies focused on insulin secretion, undoubtedly because of the
established importance of insulin secretion in the pathogenesis
of diabetes, but the findings were inconclusive (7). Insulin resis-
tance and its role in the pathogenesis of T2D were less well appre-
ciated, and, as a result, insulin sensitization was not considered as
a potential mechanism for glucose lowering at the time. It wasn’t
until much later that studies looking at a role for inflammation in
the pathogenesis of insulin resistance reinvestigated the hypogly-
cemic actions of salicylates and identified the molecular target to
be the IκB kinase-β (IKKβ)/NF-κB axis (8–10).
Although epidemiological associations relating inflammation
to T2D or obesity can be traced to the late 1950s and 1960s, when
increases were found in circulating concentrations of fibrinogen
Nonstandard abbreviations used: AGE, advanced glycation end product; CRP,
C-reactive protein; CVD, cardiovascular disease; IKKβ, IκB kinase-β; IRS-1, insulin
receptor substrate-1; MCP-1, monocyte chemoattractant protein-1; MIP, macrophage
inflammatory protein; PAI-1, plasminogen activator inhibitor-1; RAGE, receptor for
advanced glycation end products; T2D, type 2 diabetes; TZD, thiazolidinedione.
Conflict of interest: The authors have declared that no conflict of interest exists.
Citation for this article: J. Clin. Invest. 116:1793–1801 (2006). doi:10.1172/JCI29069.
and other acute-phase reactants (11–13), the findings similarly
failed to influence thoughts about pathogenesis. More recently,
additional epidemiological studies confirmed and extended these
early findings (14). Increased levels of markers and mediators
of inflammation and acute-phase reactants such as fibrinogen, 
C-reactive protein (CRP), IL-6, plasminogen activator inhibi-
tor-1 (PAI-1), sialic acid, and white cell count correlate with inci-
dent T2D (15–25). Markers of inflammation and coagulation are
reduced with intensive lifestyle intervention, as was performed
in the diabetes prevention program (26), but experiments show-
ing that adipose tissue–derived proinflammatory cytokines such
as TNF-α could actually cause insulin resistance in experimental
models provided the necessary impetus to begin thinking in terms
of pathogenesis (27–29). This discovery gave the field a critical
boost, because epidemiological studies, while highly informative,
are correlative by nature and, alone, are unable to determine cau-
sality. These different areas of research have coalesced sufficiently
that credible hypotheses can now link inflammation to the devel-
opment of insulin resistance and pathogenesis of T2D (30, 31).
Molecular pathways that link inflammation
and insulin resistance
Hotamisligil and colleagues (27) and Karasik and colleagues (28)
first showed that the proinflammatory cytokine TNF-α was able
to induce insulin resistance. This was a revolutionary idea, that
a substance produced by fat — and overproduced by expanded
fat — had local and potentially systemic effects on metabolism.
The concept of fat as a site for the production of cytokines and
other bioactive substances quickly extended beyond TNF-α to
include leptin, IL-6, resistin, monocyte chemoattractant protein-1 
(MCP-1), PAI-1, angiotensinogen, visfatin, retinol-binding pro-
tein-4, serum amyloid A (SAA), and others (32–36). Adiponectin is
similarly produced by fat, but expression decreases with increased
adiposity (37). While leptin and adiponectin are true adipokines
that appear to be produced exclusively by adipocytes, TNF-α, IL-6, 
MCP-1, visfatin, and PAI-1 are expressed as well at high levels in
activated macrophages and/or other cells. The relative amount of
each produced by the adipocyte versus associated adipose tissue
macrophages remains unknown. Sites of resistin production are
more complex; they include macrophages in humans but both
adipocytes and macrophages in rodents (34). TNF-α, IL-6, resis-
tin, and undoubtedly other pro- or antiinflammatory cytokines
appear to participate in the induction and maintenance of the
subacute inflammatory state associated with obesity. MCP-1

The Journal of Clinical Investigation      http://www.jci.org      Volume 116      Number 7      July 2006 1793
review series
and other chemokines have essential roles in the recruitment of
macrophages to adipose tissue. These cytokines and chemokines
activate intracellular pathways that promote the development of
insulin resistance and T2D.
The investigations that focused on intracellular pathways activat-
ed by inflammation, instead of individual cytokines, have helped
to restructure the framework for thinking about insulin resistance.
As mentioned above, the antihyperglycemic effects of salicylates
focused attention on IKKβ and NF-κB (38–40). However, increasing
adiposity activates both JNK and IKKβ (8, 41–43). Many of the more
typical proinflammatory stimuli simultaneously activate JNK and
IKKβ pathways, including cytokines and TLRs (Figure 1). Concor-
dantly, genetic or chemical inhibition of either JNK or IKKβ/NF-κB  of turnover. Prolonged hyperglycemia and the accompanying pro-
can improve insulin resistance. The several hypothesized mecha-
nisms that might explain how obesity activates JNK and NF-κB can
be separated into receptor and nonreceptor pathways (Figure 1).
Proinflammatory cytokines such as TNF-α and IL-1β activate JNK
and IKKβ/NF-κB through classical receptor-mediated mechanisms
that have been well characterized (Figure 1). JNK and IKKβ/NF-κB
are also activated by pattern recognition receptors, defined as sur-
1794 The Journal of Clinical Investigation      http://www.jci.org      Volume 116      Number 7      July 2006
Figure 1
Potential cellular mechanisms for activating
inflammatory signaling. Obesity and high-fat
diet activate IKKβ/NF-κB and JNK pathways
in adipocytes, hepatocytes, and associated
macrophages. Stimuli that have been shown to
activate these pathways during metabolic dys-
regulation include ligands for TNF-α, IL-1, Toll,
or AGE receptors (TNFR, IL-1R, TLR, or RAGE,
respectively), intracellular stresses including
ROS and ER stress, ceramide, and various PKC
isoforms. Obesity-induced IKKβ activation leads
to NF-κB translocation and the increased expres-
sion of numerous markers and potential media-
tors of inflammation that can cause insulin resis-
tance. Obesity-induced JNK activation promotes
the phosphorylation of IRS-1 at serine sites that
negatively regulate normal signaling through the
insulin receptor/IRS-1 axis. Examples include
serine-302 (pS302) and serine-307 (pS307). By
contrast, evidence has not been reported for obe-
sity-induced effects on transcription factors such
as AP-1 that are regulated by JNK. IKKβ and/or
NF-κB are inhibited or repressed by the actions
of salicylates, TZDs, and statins.
face proteins that recognize foreign substances. These include the
TLRs and the receptor for advanced glycation end products (RAGE)
(44). Many TLR ligands are microbial products, including LPS and
lipopeptides derived from bacteria (44). The fact that TLRs recog-
nize microbial lipid conjugates has led to speculation that endog-
enous lipids or lipid conjugates might also activate 1 or more of the
TLRs in obesity, a possibility supported by experiments showing
that saturated fatty acids bind and activate TLR4 (45). Likewise,
RAGE binds a variety of ligands, including endogenous advanced
glycation end products (AGEs) and a distinct set of microbial prod-
ucts (46, 47). AGEs are nonenzymatic adducts formed between
glucose and targeted proteins, particularly those with slow rates
duction of excess quantities of AGEs can activate NF-κB.
In addition to proinflammatory cytokine and pattern recogni-
tion receptors, cellular stresses activate JNK and NF-κB, including
ROS and ER stress. Systemic markers of oxidative stress increase
with adiposity, consistent with a role for ROS in the development
of obesity-induced insulin resistance (48). One potential mecha-
nism is through the activation of NADPH oxidase by lipid accu-
 review series

that they do so through dis-

Box 1 similar mechanisms. JNK is
NF-κB target gene products with potential involvement in the pathogenesis of insulin resistance, a stress kinase that normally
T2D, and/or atherosclerosis phosphorylates the c-Jun com-
ponent of the AP-1 transcrip-
Cytokines and Transcription Receptors and Others
tion factor, but to date there are
chemokines factors surface proteins
no known links between this
TNF-α TNFR p55 p65 RelA PAI-1 well-established transcription-
IFN-γ TNFR p75 NF-κB p50 SAA
al pathway and JNK-induced
Resistin IFNR subunits IKKa Angiotensinogen
IL-1β IL-1R IKKβ CRP insulin resistance. Instead, JNK
IL-6 IL-6R IκBα COX2 has been shown to promote
IL-8 CD40 A20 iNOS insulin resistance through the
IL-10 CD40 ligand VEGF phosphorylation of serine resi-
IL-12 E-selectin IGFBPs dues in IRS-1 (41, 42, 51, 57,
IL-18, P-selectin MnSOD 58) (Figure 1). Insulin receptor
Lymphotoxins ICAM-1
signaling that normally occurs
TGF-β VCAM-1
MCP-1 CCR2 through a tyrosine kinase cas-
MIP-1α TLR2 cade is inhibited by counter-
MIP-1β TLR4 regulatory serine/threonine
MIP-2 Lox-1 phosphorylations (59).
MIP-3α RAGE In contrast, IKKβ is highly
RANTES selective toward its physiologi-
IFNR, IFN receptor; IGFBP, IGF-binding protein; IL-1R, IL-1 receptor; MnSOD, manganese superoxide dismutase; SAA,
cal substrates, the IκB protein
serum amyloid A; TNFR, TNF receptor. inhibitors of NF-κB. Phos-
phorylation by IKKβ targets
IκBα for proteasomal degrada-
tion, which liberates NF-κB for
mulation in the adipocyte, which increases ROS production (49). translocation into the nucleus, where it promotes the expression
This mechanism was shown to increase the production of TNF-α,  of numerous target genes whose products induce insulin resis-
IL-6, and MCP-1, and decrease the production of adiponectin. tance (see Box 1). Unlike JNK, IKKβ does not phosphorylate IRS-1
Consistent with this, the antioxidant N-acetyl cysteine can reduce to cause insulin resistance but causes insulin resistance through
ROS and improve insulin resistance in a hyperglycemia-induced transcriptional activation of NF-κB. Increased lipid deposition in
model (50). Lipid accumulation also activates the unfolded pro- adipocytes leads to the production of proinflammatory cytokines,
tein response to increase ER stress in fat and liver (51). ER stress including TNF-α, IL-6, IL-1β, and resistin, which further activate
was shown to activate JNK to lead to serine phosphorylation of JNK and NF-κB pathways through a feed-forward mechanism. In
insulin receptor substrate-1 (IRS-1), but as with all of the stimuli addition to the cytokines, there is upregulated expression of tran-
described in Figure 1, ER stress similarly activates NF-κB. scriptions factors, receptors, and other relevant proteins including
Ceramides may form under conditions of cell stress and promote chemokines that recruit monocytes and stimulate their differen-
cellular signaling, including the regulation of apoptosis. Saturated tiation into macrophages (Box 1).
fats may also promote the synthesis of ceramides, which accumu- Inflammation is also closely linked to the pathogenesis of
late in tissues such as muscle and may correlate with the degree of atherosclerosis, suggesting that inflammation might be a com-
insulin resistance (52). Ceramides can also activate inflammatory mon denominator that links obesity to many of its pathological
pathways including JNK and NF-κB (53). Lipid excess increases the sequelae. Overlapping collections of transcriptionally regulated
activities of various PKC isoforms. Acute lipid infusion activates inflammatory proteins participate in the pathogenesis of these
PKC-θ in rodent muscle and PKC-βII and PKC-δ in human muscle, disorders (Box 1). Signs of inflammation accompany even the
which may also activate IKKβ and NF-κB (54, 55). Consistent with earliest accumulation of lipid within the arterial wall, including
these findings, salicylates improve muscle insulin resistance after the upregulation of the cell adhesion molecules P- and E-selec-
lipid infusion (56), although the relevance of the acute lipid infu- tin, ICAM-1, and VCAM-1, which localizes circulating immune
sion model to the pathology of obesity-induced insulin resistance is cells (60, 61). The adherent cells can migrate into the subendo-
questionable. One or more of the upstream activators of inflamma- thelial space, where they contribute to the local inflammatory
tion listed in Figure 1 may be relevant to the in vivo pathogenesis response. Cytokines and chemokines produced locally include
of insulin resistance, but more information is needed to determine MCP-1 and macrophage inflammatory protein-1α (MIP-1α),
which of these or others predominate. Once activated, however, the MIP-1β, MIP-2, and MIP-3α. T cell activation leads to expres-
processes can be self-perpetuating through a positive-feedback loop sion of IFN-γ and lymphotoxin; macrophages, endothelial cells,
created by the produced proinflammatory cytokines. and SMCs produce TNF-α (62, 63); and together these stimulate
the local production of IL-6 in the atheroma (64, 65). The addi-
Transcription versus phosphorylation tional engagement of CD40 and CD40 ligand promotes these
in the pathogenesis of insulin resistance processes as well as increasing MMP expression (66). The latter
While both JNK and IKKβ/NF-κB play important roles in inflam- can break down collagen and tissue factor, an important media-
mation-induced insulin resistance, accumulated evidence suggests tor of thrombosis (67).

The Journal of Clinical Investigation      http://www.jci.org      Volume 116      Number 7      July 2006 1795
review series

Figure 2
Potential mechanisms for activation of inflammation in adipose tissue. Dietary excess and obesity cause lipid accumulation in adipocytes, initi-
ating a state of cellular stress and activation of JNK and NF-κB. These inflammatory signaling pathways regulate protein phosphorylation and
cellular transcriptional events, thereby leading to increased adipocyte production of proinflammatory cytokines, including TNF-α, IL-6, leptin, and
resistin, chemokines such as MCP-1, and other proatherogenic mediators, for example PAI-1. Endothelial adhesion molecules (e.g., ICAM-1
and VCAM-1) and chemoattractant molecules (designated CCX) bind integrins and chemokine receptors (CCR), respectively, on the monocyte
surface to recruit them to the adipose tissues. The monocytes that differentiate into macrophages produce many of the same inflammatory
cytokines and chemokines as those listed above, and additional ones, to further promote local inflammation and propagate the inflammatory
diathesis systemically. pS, phosphoserine.

As vascular remodeling progresses, the accumulation of foam Where is inflammation-induced

cells leads to the formation of a lipid pool, rich in prothrombotic
tissue factor. PDGF and TGF-β increase the rate of collagen pro-
duction, which contributes to the formation of a fibrous cap.
Conversely, IFN-γ halts collagen synthesis by SMCs. Activated
macrophages secrete MMPs, which degrade collagen and render
the fibrous cap weak and prone to rupture (68). Thus a dynamic
balance is maintained between collagen synthesis and breakdown.
If proinflammatory forces predominate, the fibrous cap may thin
and eventually rupture, with release of prothrombotic lipids into
the lumen. This may herald the onset of an acute ischemic event.
NF-κB regulates many of the proteins that mediate the atherogenic
process, in common with the pathogenesis of insulin resistance,
suggesting that small increases in obesity-induced inflammation
might promote both processes via common mechanisms. This also
suggests the corollary that pharmacological decreases in inflam-
matory activity might coordinately downregulate the production
of a number of proteins involved in the pathogenesis of insulin
resistance, T2D, and cardiovascular disease (CVD).
insulin resistance initiated?
Adipose tissue has attracted a great deal of attention as a pathogen-
ic site of obesity-induced insulin resistance, partly because changes
in adiposity are easy to see but also because fat produces bioactive
proteins that are readily detected and reflect the inflammatory
state of the organ. However, it has also been firmly established
that all fat is not equal; adipose tissue in the subcutaneous versus
abdominal or visceral depots differs by cell size (69, 70), metabolic
activity, and potential role in insulin resistance (71, 72). Visceral fat
is more pathogenic. The adipocyte itself is integral to the develop-
ment of obesity-induced inflammation. As discussed above, pro-
teins produced by adipocytes that might collaboratively initiate the
process include TNF-α, IL-6, resistin, leptin, adiponectin, MCP-1,
PAI-1, and angiotensinogen, but since the recruited immune cells
produce many of the same substances, with the exception of leptin
and adiponectin, it is difficult to pinpoint precise sites of produc-
tion. Myeloid-selective deletion of IKKβ improves obesity-induced
insulin resistance, underscoring potential roles for both NF-κB

1796 The Journal of Clinical Investigation      http://www.jci.org      Volume 116      Number 7      July 2006

and inflammatory cells including macrophages (73), but this has
not helped to distinguish where the process is initiated. Both cell
types, lipid-laden adipocytes and recruited macrophages, seem to
participate in the pathogenesis of inflammation-induced insulin
resistance (Figure 2). Since the bulk of accumulated lipid is stored
in adipocytes, it is generally assumed that the adipocyte initiates
the process and the macrophage serves to amplify the signal. Since
many of the bioactive proteins involved are NF-κB targets, and  this has not yet been reported. In addition to cytokine and protease
NF-κB activation can be self-sustaining, antiinflammatory thera-
pies including salicylates may coordinately decrease their expres-
sion and improve insulin resistance.
It is important to understand how increasing adiposity leads to
the recruitment of immune cells to adipose tissue. MCP-1 (CCL2),
a chemoattractant for monocytes, DCs, and memory T cells, is
produced by adipocytes in parallel with increasing adiposity (74,
75), suggesting that MCP-1 might play a role in recruitment of
monocytes. Consistent with this, mice lacking CCR2, an important
receptor for MCP-1, are partly protected from developing high-fat
diet–induced insulin resistance and exhibit reductions in adipose
tissue macrophage recruitment and inflammatory gene expression
(76). The fact that protection is incomplete implies that additional
chemoattractant ligand-receptor pairs might be involved. It is also
interesting that some macrophages found in the adipose tissue of
obese rodents are large and multinucleated (74, 75, 77). Such mul-
tinucleate giant cells are often found at sites of chronic inflamma-
tion and result from the fusion or engulfment of macrophages by
each other. Macrophages including multinucleate giant cells may
aggregate at sites of adipocyte necrosis (77).
Other cell types in adipose tissue may also participate in the
inflammatory process. Vascular cells are an obvious place to look.
Adipose tissue is highly vascularized, with multiple capillaries in
contact with each adipocyte (78). Moreover, adipose tissue rap-
idly proliferates and expands as nutrient stores increase, possibly
using processes similar to the angiogenesis that supports tumor
growth (77). In addition to being important for fat expansion, the
microvasculature undoubtedly plays important roles in adipose
tissue inflammation. For example, circulating leukocytes do not
adhere to normal endothelium, but after initiation of a high-fat
Western diet the endothelium expresses cell adhesion molecules
that bind leukocytes (79). Adipose tissue endothelial cells may
increase the expression of one or more of the adhesion proteins
The Journal of Clinical Investigation      http://www.jci.org      Volume 116      Number 7      July 2006 1797
review series
Figure 3
Potential mechanisms for adiposity-induced inflam-
mation in the liver. Healthy liver contains a broad
repertoire of cells that participate in inflammatory
and immune responses, including resident hepatic
macrophages (Kupffer cells), B and T cells, NK and
NKT cells, DCs, liver sinusoidal endothelial cells,
hepatic stellate cells, and hepatocytes. Hepatic
steatosis and obesity are accompanied by the acti-
vation of inflammatory signaling pathways in liver.
Proinflammatory cytokines and FFAs, produced
either by hepatocytes in response to steatosis or
by abdominal fat tissue, may activate Kupffer cells.
Numbers of regulatory NKT cells decrease in par-
allel with the Kupffer cell activation.
ICAM-1, VCAM-1, E-selectin, or P-selectin in response to increased
adiposity. As mentioned earlier, MCP-1 induces the migration of
blood monocytes into the subendothelial space and augments
differentiation into macrophages. Thus changes are predicted in
adipose tissue endothelial cells in response to altered adiposity. As
macrophages rarely function alone, other types of immune cells
are likely to participate in adipose tissue inflammation, although
release, the actions of macrophages in inflammation include anti-
gen presentation and T cell activation. Classical inflammation
involves the added presence of neutrophils, DCs, NK cells, mast
cells, and various subtypes of T lymphocytes. Potential roles for
these other immune cells in adipose tissue inflammation will
doubtless be topics for future investigation.
In addition to adipose tissue, the liver is affected by obesity
(Figure 3). Nonalcoholic fatty liver disease (NAFLD) often accom-
panies abdominal adiposity, and its prevalence is increasing and
closely parallels the prevalence of the comorbid conditions T2D
and hyperlipidemia. The pathological spectrum of NAFLD ranges
from simple steatosis to steatohepatitis, advanced fibrosis, and
cirrhosis. Inflammation clearly plays a pivotal role in the pro-
gression of this disease process. While inadequate suppression of
hepatic glucose production, due at least in part to hepatic insulin
resistance, is an established contributor to hyperglycemia in T2D,
the role of inflammation in the pathogenesis of these processes
has only recently been explored. Inflammatory gene expression
increases in liver with increasing adiposity (43). This suggests that
hepatocyte lipid accumulation (steatosis) might induce a subacute
inflammatory response in liver that is similar to the adipose tis-
sue inflammation that follows adipocyte lipid accumulation.
Alternatively, proinflammatory substances in the portal circula-
tion, potentially produced in abdominal fat, might initiate hepatic
inflammation. Regardless, NF-κB is activated in the hepatocyte,
and cytokines including IL-6, TNF-α, and IL-1β are overproduced
in fatty liver. The proinflammatory cytokines participate in the
development of insulin resistance and activate Kupffer cells,
the resident hepatic macrophages. Unlike adipose tissue, where
macrophages are relatively sparse basally but increase numerically
with adiposity, the liver is densely populated with Kupffer cells,
which account for over 5% of total cells. The number of Kupffer
cells does not increase with adiposity, but their activation state
review series

Figure 4
Local, portal, and systemic effects of inflammation in insulin resistance and atherogenesis. Increasing adiposity activates inflammatory responses
in fat and liver, with associated increases in the production of cytokines and chemokines. Immune cells including monocytes and macrophages
are recruited and/or activated, and together these cause local insulin resistance. Portal delivery of abdominal fat–derived cytokines and lipids
contributes to hepatic inflammation and insulin resistance. Proinflammatory and proatherogenic mediators are produced in the adipose tissue
and liver and associated immune cells. This creates a systemic inflammatory diathesis that promotes insulin resistance in skeletal muscle and
other tissues and atherogenesis in the vasculature.

does (43). A wide variety of other immune cells are present in nor-
mal liver and may also play roles in inflammation-induced insulin
resistance, including T and B lymphocytes, NK cells, and DCs as
well as hepatic stellate cells and liver sinusoidal endothelial cells
(80). NKT cells are enriched in normal mouse liver, and their num-
bers decrease in ob/ob or high-fat diet models of obesity (81, 82).
NKT cells are regulatory lymphocytes, with features of both clas-
sical T (CD3+) and NK (NK1.1+) cells and characteristic expression
of CD1d, an MHC class I homologue that presents glycolipid anti-
gens to TCRs. Adoptive transfer of NKT cells reportedly improves
nonalcoholic steatohepatitis and glucose intolerance in ob/ob
mice (83), consistent with a role for decreased NKT numbers in
the pathogenesis of these disorders.
Skeletal muscle is another major site of insulin resistance in obe-
sity and T2D. However, increasing adiposity does not appear to
activate inflammatory cascades in skeletal muscle, as it does in fat
and liver. Inflammation is activated in muscle by intralipid infu-
sion, but this is distinct from the effects of increasing adiposity.
The lipid infusion model is a research tool used to acutely raise
circulating and intratissue fatty acid levels and induce insulin
resistance. Intralipid infusion activates PKC-θ and IKKβ in mouse
muscle, and the associated insulin resistance is inhibited by salicy-
late or IKKβ depletion (54, 56). This is in contrast with high-fat
diet–induced and obesity-induced insulin resistance, neither of
which activates IKKβ/NF-κB in skeletal muscle (84, 85) or leads
to an increase in skeletal muscle macrophages (75). Concordantly,
neither muscle-specific ablation of IKKβ nor muscle-specific inhi-
bition of NF-κB improves insulin resistance in obese mice (84, 85).
It is perhaps more appropriate to think of skeletal muscle as a tar-
get of inflammation-induced insulin resistance as opposed to a
site of initiation (Figure 4).
Future trials to target inflammation
While demonstrating that TNF-α could induce insulin resistance
(27, 28) was paramount in linking inflammation to the pathogene-
sis of insulin resistance, studies to evaluate whether blocking TNF-α 
with mAbs improves insulin sensitivity have so far been inconclu-
sive (86, 87). This may be due in part to small trial sizes and inad-
equate statistical power but might also be related to mechanisms of
pathogenesis. TNF-α by itself may just not be that important. With
multiple cytokines and chemokines involved, it is also possible that
different combinations of cytokines and chemokines might be
more or less important in different patients. If this were true, then
neutralization strategies might need to be individually tailored. It
may be more prudent to modulate signaling at convergent sites of
integration, such as JNK or IKKβ/NF-κB, since this would provide
a more general strategy for treatment. Several drugs in current clini-
cal practice have been shown to have antiinflammatory properties
or side effects distinct from their major mechanisms of action,
including members of the thiazolidinedione (TZD) class of PPARγ
agonists and members of the statin class of HMG CoA reductase
inhibitors. Both appear to have important antiinflammatory prop-

1798 The Journal of Clinical Investigation      http://www.jci.org      Volume 116      Number 7      July 2006

erties and potential benefits beyond their primary actions on glu-
cose homeostasis and cholesterol lowering, respectively.
TZDs, which include pioglitazone, rosiglitazone, and trogli-
tazone (the latter is no longer in clinical use), are used to improve
insulin sensitivity and reduce hyperglycemia in patients with T2D.
The TZDs’ primary mode of action is through binding and activat-
ing PPARγ to induce the expression of a number of gene products
with roles in adipocyte differentiation, lipid and glucose uptake,
and fatty acid storage (88). TZDs’ beneficial actions in insulin sen-
sitivity are often attributed to fatty acid sequestration in the adi-
pose tissue. This reduces circulation levels of FFAs and keeps the
fatty acids out of other tissues, including muscle and liver, where
they can cause insulin resistance. But in addition to adipocytes,
PPARγ is also expressed in macrophages and other immune cells,
hepatocytes, endothelial cells, and VSMCs. The antiinflammatory
actions of the TZDs appear to be mediated through the transre-
pression of NF-κB and consequent decreases in the expression of
target genes for cytokine and growth factors, cell proliferation and
migration, ECM remodeling, and cell cycle progression and dif-
ferentiation. One proposed mechanism involves the TZD-depen-
dent SUMOylation of PPARγ, which targets it to nuclear receptor
corepressor–histone deacetylase-3 (NCoR-HDAC3) complexes
on inflammatory gene promoters (such as NF-κB and AP-1). The
binding of SUMOylated PPARγ stabilizes the corepressor complex-
es and prevents their clearance, thus maintaining the transcription
factor complexes in a repressed state (89). It has even been sug-
gested that TZDs exert antiinflammatory effects through PPARγ-
independent activation of the glucocorticoid receptor (90). The
clinical benefits of TZDs may in part depend upon antiinflam-
matory effects that work together with classical mechanisms of
glucose and lipid regulation (91) to improve insulin sensitivity,
promote plaque remodeling, and potentially reduce cardiovascular
events (92, 93). Several other nuclear receptors, including PPARα,
PPARδ, and liver X receptor (LXR), have also been shown to have
antiinflammatory properties that may be beneficial for the treat-
ment of metabolic or cardiovascular conditions (94, 95), although
significant glucose-lowering effects have not been associated with
agonists to these transcription factors.
Several of the clinically available statins have been shown to down-
regulate transcriptional activities of NF-κB, AP-1, and HIF-1α,  syndromes in patients with lipodystrophy who lack adipose tis-
with coordinate reductions in the expression of prothrombotic and
inflammatory cytokines (95). Randomized clinical trials evaluat-
ing statins have also demonstrated reductions in CRP, multiple
cytokines, and inflammatory markers. Despite these modest anti-
inflammatory properties, the statins do not appear to significantly
influence either insulin resistance or glycemia.
The antiinflammatory properties of TZDs and statins are side
effects distinct from their primary modes of action. By contrast,
high-dose salicylates directly target inflammation by inhibiting
NF-κB (38–40). The glucose-lowering effects of salicylates, seen
decades ago in patients with diabetes (1–7), are now recognized as
well to be due at least in part to NF-κB inhibition (8–10, 43, 73).
More recent studies demonstrated that high-dose aspirin (∼7.0 g/d)  insulin resistance and T2D and modulate risk for CVD and other
improved multiple metabolic measures in patients with T2D,
including substantial reductions in fasting and postprandial glu-
cose, triglycerides, and FFAs (9). These changes were associated
with reduced hepatic glucose production and improvements in
insulin-stimulated glucose disposal, assessed during euglycemic-
hyperinsulinemic clamping. However, the therapeutic usefulness of
high-dose aspirin is limited by the antithrombotic and anti–platelet
The Journal of Clinical Investigation      http://www.jci.org      Volume 116      Number 7      July 2006 1799
review series
aggregation effects, which, coupled with gastrointestinal irritation,
are associated with unacceptably high risks of bleeding.
Aspirin is a prototypical NSAID that effectively inhibits the
COX enzymes COX1 and COX2 through the irreversible modifi-
cation of the enzymes’ active sites. This occurs through a trans-
acetylation reaction; the acetyl group of aspirin migrates to and
covalently modifies the enzymes to block their catalytic activities.
In platelets this occurs even at low doses (80–100 mg/d), because
they lack nuclei and cannot resynthesize COX1, and platelets
are thus inactivated for their lifetimes. Aspirin at a typical dose
of 650 mg covalently modifies COX1 and COX2 in all tissues,
which similarly inhibits prostaglandin synthesis. NSAIDs other
than aspirin do not transacetylate the COX enzymes but bind
with sufficiently high affinity that the enzymes are competitively
inhibited. The mechanism of action of high-dose salicylate is dis-
tinct, as it lacks an acetyl group and neither covalently modifies
COX1 or COX2 nor binds with sufficient affinity to effectively
inhibit the enzymes.
The nonacetylated salicylates, delivered as sodium salicylate, sal-
salate, and Trilisate, inhibit NF-κB (38–40). This is presumed to be
through direct inhibition of IKKβ (40). The nonacetylated salicy-
lates do not prolong bleeding times and may thus provide a rela-
tively safe and effective means of targeting the subacute inflamma-
tion that underlies the obesity-related syndromes. In fact, clinical
trials to test efficacy, tolerability, and durability of salsalate (Disal-
cid) are currently being undertaken in the NIH-funded TINSAL-
T2D and -CVD trials (Targeting Inflammation with Salsalate in
T2D or CVD, respectively). Other antiinflammatory strategies are
also being considered, including the potential use of specific JNK
or IKKβ inhibitors and compounds that block TNF-α, IL-6, TLR,
or chemokine signaling or reduce oxidative or ER stress, but these
studies are at much earlier stages.
Concluding remarks
While the importance of inflammation-induced insulin resistance
is doubtless increasing in parallel with the epidemic of obesity,
there are additional unrelated mechanisms associated with insulin
resistance. For example, polymorphisms in genes encoding com-
ponents of the insulin signaling pathway and the insulin-resistant
sue are different from the insulin resistance in typical overweight
patients with T2D. Whether antiinflammatory strategies are ben-
eficial in these cases has not been investigated. Similarly unknown
are the distinctions between populations, such as those of Asian
descent who may exhibit the characteristics of adiposity-induced
inflammation while relatively lean.
In summary, obesity, T2D, and CVD share a metabolic milieu
characterized by insulin resistance and chronic subacute inflam-
mation. While drugs that secondarily alter the inflammatory pro-
cess are undoubtedly of great clinical importance, several lines of
evidence suggest it might also be possible to directly target inflam-
mation with pharmacological interventions to treat and/or prevent
metabolic conditions. These approaches may provide clinical ben-
efits to a large number of persons affected by the obesity epidemic
and the related cluster of metabolic disorders.
Address correspondence to: Steven E. Shoelson, Joslin Diabetes
Center, One Joslin Place, Boston, Massachusetts 02215, USA. 
E-mail: steven.shoelson@joslin.harvard.edu.
review series
1. Ebstein, W. 1876. Zur therapie des Diabetes mel-
litus, insbesondere über die Anwendung des sali-
cylsauren Natron bei demselben. Berliner Klinische
Wochenschrift. 13:337–340.
2. Williamson, R.T. 1901. On the treatment of glycos-
uria and diabetes mellitus with sodium salicylate.
Br. Med. J. 1:760–762.
3. Shoelson, S. 2002. Invited comment on W. Ebstein:
on the therapy of diabetes mellitus, in particular
on the application of sodium salicylate. J. Mol. Med.
80:618–619.
4. Reid, J., Macdougall, A.I., and Andrews, M.M. 1957.
On the efficacy of salicylate in treating diabetes
mellitus. Br. Med. J. 2:1071–1074.
5. Hecht, A., and Goldner, M.G. 1959. Reappraisal of
the hypoglycemic action of acetylsalicylate. Metabo-
lism. 8:418–428.
6. Gilgore, S.G. 1960. The influence of salicylate on
hyperglycemia. Diabetes. 9:392–393.
7. Baron, S.H. 1982. Salicylates as hypoglycemic
agents. Diabetes Care. 5:64–71.
8. Yuan, M., et al. 2001. Reversal of obesity- and diet-
induced insulin resistance with salicylates or tar-
geted disruption of Ikkβ. Science. 293:1673–1677.
9. Hundal, R.S., et al. 2002. Mechanism by which
high-dose aspirin improves glucose metabolism
in type 2 diabetes. J. Clin. Invest. 109:1321–1326.
doi:10.1172/JCI200214955.
10. Shoelson, S.E., Lee, J., and Yuan, M. 2003. Inflam-
mation and the IKKβ/IκB/NF-κB axis in obesity-
and diet-induced insulin resistance. Int. J. Obes.
Relat. Metab. Disord. 27(Suppl. 3):S49–S52.
11. Fearnley, G.R., Vincent, C.T., and Chakrabarti, R.
1959. Reduction of blood fibrinolytic activity in
diabetes mellitus by insulin. Lancet. 2:1067.
12. Ogston, D., and McAndrew, G.M. 1964. Fibrinoly-
sis in obesity. Lancet. 14:1205–1207.
13. Grace, C.S., and Goldrick, R.B. 1968. Fibrinolysis
and body build. Interrelationships between blood
fibrinolysis, body composition and parameters of
lipid and carbohydrate metabolism. J. Atheroscler.
Res. 8:705–719.
14. Tataranni, P.A., and Ortega, E. 2005. A burning
question: does an adipokine-induced activation of
the immune system mediate the effect of overnu-
trition on type 2 diabetes? Diabetes. 54:917–927.
15. Schmidt, M.I., et al. 1999. Markers of inflamma-
tion and prediction of diabetes mellitus in adults
(Atherosclerosis Risk in Communities study): a
cohort study. Lancet. 353:1649–1652.
16. Duncan, B.B., et al. 2003. Low-grade systemic
inflammation and the development of type 2 dia-
betes: the Atherosclerosis Risk in Communities
study. Diabetes. 52:1799–1805.
17. Pradhan, A.D., Manson, J.E., Rifai, N., Buring, J.E.,
and Ridker, P.M. 2001. C-reactive protein, interleu-
kin 6, and risk of developing type 2 diabetes mel-
litus. JAMA. 286:327–334.
18. Barzilay, J.I., et al. 2001. The relation of markers
of inflammation to the development of glucose
disorders in the elderly: the Cardiovascular Health
Study. Diabetes. 50:2384–2389.
19. Vozarova, B., et al. 2002. High white blood cell
count is associated with a worsening of insulin
sensitivity and predicts the development of type 2
diabetes. Diabetes. 51:455–461.
20. Festa, A., D’Agostino, R., Jr., Tracy, R.P., and Haff-
ner, S.M. 2002. Elevated levels of acute-phase
proteins and plasminogen activator inhibitor-1
predict the development of type 2 diabetes: the
insulin resistance atherosclerosis study. Diabetes.
51:1131–1137.
21. Freeman, D.J., et al. 2002. C-reactive protein is an
independent predictor of risk for the development
of diabetes in the West of Scotland Coronary Pre-
vention Study. Diabetes. 51:1596–1600.
22. Ford, E.S. 2002. Leukocyte count, erythrocyte
sedimentation rate, and diabetes incidence in a
1800 The Journal of Clinical Investigation      http://www.jci.org      Volume 116      Number 7      July 2006
national sample of US adults. Am. J. Epidemiol.
155:57–64.
23. Prince, R.L., Larkins, R.G., and Alford, F.P. 1981.
The effect of acetylsalicylic acid on plasma glucose
and the response of glucose regulatory hormones
to intravenous glucose and arginine in insulin
treated diabetics and normal subjects. Metabolism.
30:293–298.
24. Nakanishi, N., Yoshida, H., Matsuo, Y., Suzuki, K.,
and Tatara, K. 2002. White blood-cell count and
the risk of impaired fasting glucose or type II dia-
betes in middle-aged Japanese men. Diabetologia.
45:42–48.
25. Spranger, J., et al. 2003. Inflammatory cytokines
and the risk to develop type 2 diabetes: results of
the prospective population-based European Pro-
spective Investigation into Cancer and Nutrition
(EPIC)-Potsdam Study. Diabetes. 52:812–817.
26. Haffner, S., et al. 2005. Intensive lifestyle interven-
tion or metformin on inflammation and coagula-
tion in participants with impaired glucose toler-
ance. Diabetes. 54:1566–1572.
27. Hotamisligil, G.S., Shargill, N.S., and Spiegelman,
B.M. 1993. Adipose expression of tumor necrosis
factor-α: direct role in obesity-linked insulin resis-
tance. Science. 259:87–91.
28. Feinstein, R., Kanety, H., Papa, M.Z., Lunenfeld, B.,
and Karasik, A. 1993. Tumor necrosis factor-α sup-
presses insulin-induced tyrosine phosphorylation
of insulin receptor and its substrates. J. Biol. Chem.
268:26055–26058.
29. Hotamisligil, G.S., and Spiegelman, B.M. 1994.
Tumor necrosis factor alpha: a key component of
the obesity-diabetes link. Diabetes. 43:1271–1278.
30. Pickup, J.C., and Crook, M.A. 1998. Is type II dia-
betes mellitus a disease of the innate immune sys-
tem? Diabetologia. 41:1241–1248.
31. Kolb, H., and Mandrup-Poulsen, T. 2005. An
immune origin of type 2 diabetes? Diabetologia.
48:1038–1050.
32. Zhang, Y., Proenca, R., Maffei, M., Barone, M.,
Leopold, L., and Friedman, J.M. 1994. Positional
cloning of the mouse obese gene and its human
homologue. Nature. 372:425–432.
33. Fried, S.K., Bunkin, D.A., and Greenberg, A.S.
1998. Omental and subcutaneous adipose tissues
of obese subjects release interleukin-6: depot dif-
ference and regulation by glucocorticoid. J. Clin.
Endocrinol. Metab. 83:847–850.
34. Steppan, C.M., et al. 2001. The hormone resistin
links obesity to diabetes. Nature. 409:307–312.
35. Shimomura, I., et al. 1996. Enhanced expression of
PAI-1 in visceral fat: possible contributor to vascu-
lar disease in obesity. Nat. Med. 2:800–803.
36. Fukuhara, A., et al. 2005. Visfatin: a protein secret-
ed by visceral fat that mimics the effects of insulin.
Science. 307:426–430.
37. Scherer, P.E., Williams, S., Fogliano, M., Baldini,
G., and Lodish, H.F. 1995. A novel serum protein
similar to C1q, produced exclusively in adipocytes.
J. Biol. Chem. 270:26746–26749.
38. Kopp, E., and Ghosh, S. 1994. Inhibition of NF-
kappa B by sodium salicylate and aspirin. Science.
265:956–959.
39. Pierce, J.W., Read, M.A., Ding, H., Luscinskas, F.W.,
and Collins, T. 1996. Salicylates inhibit I kappa
B-alpha phosphorylation, endothelial-leukocyte
adhesion molecule expression, and neutrophil
transmigration. J. Immunol. 156:3961–3969.
40. Yin, M.J., Yamamoto, Y., and Gaynor, R.B. 1998.
The anti-inflammatory agents aspirin and salicy-
late inhibit the activity of IκB kinase-β. Nature.
396:77–80.
41. Aguirre, V., Uchida, T., Yenush, L., Davis, R., and
White, M.F. 2000. The c-Jun N-terminal kinase
promotes insulin resistance during association
with insulin receptor substrate-1 and phosphory-
lation of Ser(307). J. Biol. Chem. 275:9047–9054.
42. Hirosumi, J., et al. 2002. A central role for JNK in
obesity and insulin resistance. Nature. 420:333–336.
43. Cai, D., et al. 2005. Local and systemic insulin resis-
tance resulting from hepatic activation of IKKβ
and NF-κB. Nat. Med. 11:183–190.
44. Akira, S., Uematsu, S., and Takeuchi, O. 2006.
Pathogen recognition and innate immunity. Cell.
124:783–801.
45. Lee, J.Y., Sohn, K.H., Rhee, S.H., and Hwang, D.
2001. Saturated fatty acids, but not unsaturated
fatty acids, induce the expression of cyclooxygenase-
2 mediated through Toll-like receptor 4. J. Biol.
Chem. 276:16683–16689.
46. Ramasamy, R., Yan, S.F., and Schmidt, A.M. 2005.
The RAGE axis and endothelial dysfunction:
maladaptive roles in the diabetic vasculature and
beyond. Trends Cardiovasc. Med. 15:237–243.
47. Bierhaus, A., et al. 2005. Understanding RAGE,
the receptor for advanced glycation end products. 
J. Mol. Med. 83:876–886.
48. Keaney, J.F., Jr., et al. 2003. Obesity and systemic
oxidative stress: clinical correlates of oxidative
stress in the Framingham Study. Arterioscler.
Thromb. Vasc. Biol. 23:434–439.
49. Furukawa, S., et al. 2004. Increased oxidative
stress in obesity and its impact on metabolic syn-
drome. J. Clin. Invest. 114:1752–1761. doi:10.1172/
JCI200421625.
50. Lin, Y., et al. 2005. The hyperglycemia-induced
inflammatory response in adipocytes: the role of
reactive oxygen species. J. Biol. Chem. 280:4617–4626.
51. Ozcan, U., et al. 2004. Endoplasmic reticulum
stress links obesity, insulin action, and type 2 dia-
betes. Science. 306:457–461.
52. Straczkowski, M., et al. 2004. Relationship between
insulin sensitivity and sphingomyelin signal-
ing pathway in human skeletal muscle. Diabetes.
53:1215–1221.
53. Summers, S.A. 2006. Ceramides in insulin resis-
tance and lipotoxicity. Prog. Lipid Res. 45:42–72.
54. Griffin, M.E., et al. 1999. Free fatty acid-induced
insulin resistance is associated with activation of
protein kinase C theta and alterations in the insu-
lin signaling cascade. Diabetes. 48:1270–1274.
55. Itani, S.I., Ruderman, N.B., Schmieder, F., and
Boden, G. 2002. Lipid-induced insulin resistance
in human muscle is associated with changes in dia-
cylglycerol, protein kinase C, and IkappaB-alpha.
Diabetes. 51:2005–2011.
56. Kim, J.K., et al. 2001. Prevention of fat-induced
insulin resistance by salicylate. J. Clin. Invest.
108:437–446.
57. Aguirre, V., et al. 2002. Phosphorylation of Ser307
in insulin receptor substrate-1 blocks interactions
with the insulin receptor and inhibits insulin
action. J. Biol. Chem. 277:1531–1537.
58. Werner, E.D., Lee, J., Hansen, L., Yuan, M., and
Shoelson, S.E. 2004. Insulin resistance due to
phosphorylation of insulin receptor substrate-1 at
serine 302. J. Biol. Chem. 279:35298–35305.
59. Zick, Y. 2005. Ser/Thr phosphorylation of IRS pro-
teins: a molecular basis for insulin resistance. Sci.
STKE. 268:e4.
60. Tedder, T.F., Steeber, D.A., Chen, A., and Engel, P.
1995. The selectins: vascular adhesion molecules.
FASEB J. 9:866–873.
61. Adams, D.H., and Shaw, S. 1994. Leucocyte-
endothelial interactions and regulation of leuco-
cyte migration. Lancet. 343:831–836.
62. Warner, S.J., and Libby, P. 1989. Human vascu-
lar smooth muscle cells. Target for and source of
tumor necrosis factor. J. Immunol. 142:100–109.
63. Barath, P., et al. 1990. Detection and localization
of tumor necrosis factor in human atheroma. Am.
J. Cardiol. 65:297–302.
64. Seino, Y., et al. 1994. Interleukin 6 gene transcripts
are expressed in human atherosclerotic lesions.
Cytokine. 6:87–91.

65. Rus, H.G., Vlaicu, R., and Niculescu, F. 1996.
Interleukin-6 and interleukin-8 protein and gene
expression in human arterial atherosclerotic wall.
Atherosclerosis. 127:263–271.
66. Schonbeck, U., and Libby, P. 2001. CD40 signaling
and plaque instability. Circ. Res. 89:1092–1103.
67. Mach, F., Schonbeck, U., Bonnefoy, J.Y., Pober, J.S.,
and Libby, P. 1997. Activation of monocyte/mac-
rophage functions related to acute atheroma com-
plication by ligation of CD40: induction of colla-
genase, stromelysin, and tissue factor. Circulation.
96:396–399.
68. Libby, P. 2001. Current concepts of the pathogen-
esis of the acute coronary syndromes. Circulation.
104:365–372.
69. Johnson, P.R., and Hirsch, J. 1972. Cellularity of
adipose depots in six strains of genetically obese
mice. J. Lipid Res. 13:2–11.
70. Krotkiewski, M., Bjorntorp, P., Sjostrom, L., and
Smith, U. 1983. Impact of obesity on metabolism
in men and women. Importance of regional adipose
tissue distribution. J. Clin. Invest. 72:1150–1162.
71. Coon, P.J., Rogus, E.M., Drinkwater, D., Muller,
D.C., and Goldberg, A.P. 1992. Role of body fat dis-
tribution in the decline in insulin sensitivity and
glucose tolerance with age. J. Clin. Endocrinol. Metab.
75:1125–1132.
72. Gastaldelli, A., et al. 2002. Metabolic effects of vis-
ceral fat accumulation in type 2 diabetes. J. Clin.
Endocrinol. Metab. 87:5098–5103.
73. Arkan, M.C., et al. 2005. IKK-beta links inflamma-
tion to obesity-induced insulin resistance. Nat. Med.
11:191–198.
74. Sartipy, P., and Loskutoff, D.J. 2003. Monocyte
chemoattractant protein 1 in obesity and insulin resis-
tance. Proc. Natl. Acad. Sci. U. S. A. 100:7265–7270.
75. Weisberg, S.P., et al. 2003. Obesity is associated with
macrophage accumulation in adipose tissue. J. Clin.
The Journal of Clinical Investigation      http://www.jci.org      Volume 116      Number 7      July 2006 1801
Invest. 112:1796–1808. doi:10.1172/JCI200319246.
76. Weisberg, S.P., et al. 2006. CCR2 modulates inflam-
matory and metabolic effects of high-fat feeding.  87. Gonzalez-Gay, M.A., et al. 2006. Anti-tumor necro-
J. Clin. Invest. 116:115–124. doi:10.1172/JCI24335.
77. Kolonin, M.G., Saha, P.K., Chan, L., Pasqualini, R.,
and Arap, W. 2004. Reversal of obesity by targeted
ablation of adipose tissue. Nat. Med. 10:625–632.
78. Crandall, D.L., Hausman, G.J., and Kral, J.G. 1997.
A review of the microcirculation of adipose tissue:
anatomic, metabolic, and angiogenic perspectives.
Microcirculation. 4:211–232.
79. Blake, G.J., and Ridker, P.M. 2002. Inflammatory
bio-markers and cardiovascular risk prediction.  90. Ialenti, A., et al. 2005. Mechanism of the anti-
J. Intern. Med. 252:283–294.
80. Racanelli, V., and Rehermann, B. 2006. The liver as
an immunological organ. Hepatology. 43(Suppl. 1):
S54–S62.
81. Guebre-Xabier, M., et al. 2000. Altered hepatic lym-
phocyte subpopulations in obesity-related murine
fatty livers: potential mechanism for sensitization
to liver damage. Hepatology. 31:633–640.
82. Li, Z., Soloski, M.J., and Diehl, A.M. 2005. Dietary
factors alter hepatic innate immune system in mice
with nonalcoholic fatty liver disease. Hepatology.
42:880–885.
83. Elinav, E., et al. 2006. Adoptive transfer of regula-
tory NKT lymphocytes ameliorates non-alcoholic
steatohepatitis and glucose intolerance in ob/ob
mice and is associated with intrahepatic CD8 trap-
ping. J. Pathol. 209:121–128.
84. Rohl, M., et al. 2004. Conditional disruption of IκB
kinase 2 fails to prevent obesity-induced insulin
resistance. J. Clin. Invest. 113:474–481. doi:10.1172/
JCI200418712.
85. Cai, D., et al. 2004. IKKβ/NF-κB activation causes
severe muscle wasting in mice. Cell. 119:285–298.
86. Dominguez, H., et al. 2005. Metabolic and vascu-
lar effects of tumor necrosis factor-alpha blockade
review series
with etanercept in obese patients with type 2 diabe-
tes. J. Vasc. Res. 42:517–525.
sis factor-alpha blockade improves insulin resis-
tance in patients with rheumatoid arthritis. Clin.
Exp. Rheumatol. 24:83–86.
88. Yki-Jarvinen, H. 2004. Thiazolidinediones. N. Engl.
J. Med. 351:1106–1118.
89. Pascual, G., et al. 2005. A SUMOylation-depen-
dent pathway mediates transrepression of inflam-
matory response genes by PPAR-gamma. Nature.
437:759–763.
inflammatory effect of thiazolidinediones: rela-
tionship with the glucocorticoid pathway. Mol.
Pharmacol. 67:1620–1628.
91. Xiang, A.H., et al. 2006. Effect of pioglitazone on
pancreatic beta-cell function and diabetes risk in
Hispanic women with prior gestational diabetes.
Diabetes. 55:517–522.
92. Meisner, F., et al. 2006. Effect of rosiglitazone treat-
ment on plaque inflammation and collagen con-
tent in nondiabetic patients: data from a random-
ized placebo-controlled trial. Arterioscler. Thromb.
Vasc. Biol. 26:845–850.
93. Dormandy, J.A., et al. 2005. Secondary preven-
tion of macrovascular events in patients with type
2 diabetes in the PROactive Study (PROspective
pioglitAzone Clinical Trial In macroVascular
Events): a randomised controlled trial. Lancet.
366:1279–1289.
94. Glass, C.K., and Ogawa, S. 2006. Combinato-
rial roles of nuclear receptors in inflammation and
immunity. Nat. Rev. Immunol. 6:44–55.
95. Castrillo, A., and Tontonoz, P. 2004. Nuclear recep-
tors in macrophage biology: at the crossroads of
lipid metabolism and inflammation. Annu. Rev.
Cell Dev. Biol. 20:455–480.